Uma aplicação de modelos baseados em agentes para avaliação prospectiva de intervenções em saúde

Abstract

Simulation models have become relevant tools in public health, as they allow for the prospective evaluation of interventions in complex scenarios with lower costs and greater flexibility compared to traditional methods. Among these approaches, agent-based models stand out for their ability to represent populations at the individual level, accounting for heterogeneity and interactions capable of producing emergent population-level effects. This study revisits the model proposed by Bellido (2017), which investigated the impact of interventions aimed at reducing Body Mass Index (BMI) on the incidence of Systemic Arterial Hypertension (SAH). The original model was expanded through the incorporation of a social network structure among agents and the introduction of heterogeneous adherence probabilities to interventions. The results showed that interventions based on physical activity and calorie-restricted diets reduced the incidence of SAH, even when implemented at later stages of life. The different network topologies used revealed variations in the ability of interventions to influence outcomes. The inclusion of a social network structure brought the model closer to real-world population behavior by incorporating an important mechanism related to the development of SAH, increasing its predictive ability on the effectiveness of interventions and showing the viability of this kind of study to generate experimental evidence.